---

license: apache-2.0

base_model:

• TokenRhythm/NeoHorse-1-4B

pipeline_tag: text-generation

language:

• en

library_name: transformers

tags:

• text-generation-inference
• llama-cpp
• agentic
• tool-use
• coding
• reasoning

- instruction-following

---

NeoHorse-1-4B-GGUF

> NeoHorse-1-4B is a 4-billion-parameter causal language model from TokenRhythm, post-trained from Qwen3.5-4B as an initial prototype on the path toward recursive self-improvement (RSI), targeting text-based agent harnesses, tool use, coding, and instruction following. Like its 9B sibling, its core innovation is a routing harness that assigns tasks to a heterogeneous model pool, records tool interactions and outcomes, estimates capability demand, and feeds that signal back into shaping the next training mixture via routing-guided curriculum SFT and on-policy distillation, backed by rigorous deduplication, decontamination, and six-dimensional semantic data evaluation. This release contains language-model weights only (vision weights excluded, repackaged for text-only inference) and retains a 262,144-token native context window extensible to 1,010,000. Across a ten-benchmark evaluation against Qwen3.5-4B, Gemma-4-E4B-it, Nanbeige-4.2-3B, Agents-A1-4B, and Spark-X2.5-4B, NeoHorse-1-4B posts the best overall macro average (64.87 vs. 58.94 for its Qwen3.5-4B base, a +5.93 gain), with the largest improvements on agentic benchmarks like VitaBench (+10.50), WorkBuddy Bench (+9.79), HumanEval (+9.75), and QwenClawBench (+6.21), alongside a strong tau2-Bench score of 88.46 (best in the comparison set). It's servable via SGLang or vLLM with Qwen3-style reasoning and tool-call parsers, and is released under the Apache License 2.0.

Model Files

File Name | Quant Type | File Size | File Link |

|-----------|------------|-----------|-----------|

| NeoHorse-1-4B.BF16.gguf | BF16 | 8.42 GB | Download |

| NeoHorse-1-4B.Q3_K_L.gguf | Q3_K_L | 2.42 GB | Download |

| NeoHorse-1-4B.Q3_K_M.gguf | Q3_K_M | 2.26 GB | Download |

| NeoHorse-1-4B.Q3_K_S.gguf | Q3_K_S | 2.07 GB | Download |

| NeoHorse-1-4B.Q4_0.gguf | Q4_0 | 2.54 GB | Download |

| NeoHorse-1-4B.Q4_K_M.gguf | Q4_K_M | 2.71 GB | Download |

| NeoHorse-1-4B.Q4_K_S.gguf | Q4_K_S | 2.56 GB | Download |

| NeoHorse-1-4B.Q5_0.gguf | Q5_0 | 2.99 GB | Download |

| NeoHorse-1-4B.Q5_K_M.gguf | Q5_K_M | 3.07 GB | Download |

| NeoHorse-1-4B.Q5_K_S.gguf | Q5_K_S | 2.99 GB | Download |

llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp